United Airlines Flight 85 (UAL85) encountered severe turbulence as it was on approach to landing at Newark Liberty International Airport (KEWR) — which injured 22 passengers, with 7 of those being transported to a hospital (media report 1 | media report 2) late in the day on 29 March 2024. Gusty NW winds may have contributed to the turbulence (which likely occurred around 2138 UTC, according to flight data from FlightAware — note the rapid rate of ascent beginning at that time); the peak wind gust at KEWR was 48 knots or 55 mph at 2039 UTC. After UAL85 aborted its landing at KEWR, it diverted north to New York Stewart International Airport (KSWF), landing at 2208 UTC (above).

1-minute Mesoscale Domain Sector GOES-16 (GOES-East) Low-level Water Vapor (7.3 µm) images (below) depicted widespread mountain waves across southeastern New York and northern New Jersey during that time. These mountain waves could also have played a role in creating boundary layer turbulence. The Pilot Report (PIREP) for this UAL85 incident did not appear in the real-time AWIPS data, probably due to time restraints imposed upon the flight crew to get the aircraft safely diverted as quickly as possible.

In a faster animation of GOES-16 Water Vapor images without plots of surface reports (below), note that the mountain wave structure across northern New Jersey (including near KEWR) becomes slightly deformed, and even begins to move westward a few miles.

A toggle between GOES-16 Water Vapor imagery and Topography (below) indicated that these mountain waves were forming downwind of the Catskills in New York and the Poconos in Pennsylvania.

Plots of rawinsonde data from Albany NY and Upton NY (below) both showed adiabatic or near-adiabatic lapse rates below the 800-850 hPa pressure level — which aided in the downward transport of momentum (stronger winds) aloft to the surface earlier in the day.

View only this post Read Less

NOAA-20’s early morning orbit and clear skies allowed a distinct view of Sea Surface Temperatures over the Gulf of Mexico and the East Coast on 29 March 2024, as shown above (The data, derived from the CIMSS Direct Broadcast Antenna) are available via an LDM feed from CIMSS). The Loop Current in the Gulf of Mexico (with surface temperatures from 80-81^oF) is obvious, as is the Gulf Stream along the East Coast.

How did the VIIRS Sea Surface Temperature values compare with co-located Fixed Buoy values? In the 3 examples below, the agreement was quite good — within 1ºF. Buoy SST values appear in the lower right corner of their station model plots.

Two aspects (at least!) of this imagery bear mention. The Mississippi River delta (enlarged, below) is a source of cool water flowing into the Gulf. The cyan values are temperatures from 58-61^oF vs low 70s ^oF (in yellow) just offshore.

There is a very tight thermal gradient northeast of the Outer Banks of North Carolina! The magenta color shows temperatures in the upper 40s oF vs upper 70s (in orange) less than 20 miles away!

---

How do the VIIRS SSTs compare to GOES-16’s? The toggle below over the northern Gulf shows differences related to cloud detection over the cooler waters of the Mississippi River delta, and along the edge of the warmer waters offshore: NOAA-20 has a more complete picture.

The NOAA-20 imagery is also more complete near the Outer Banks as shown in the toggle below.

View only this post Read Less

LightningCast is a product within the ProbSevere portfolio that uses 4 ABI (or AHI) channels to estimate the likelihood of a GLM observation within the next 60 minutes. For the past couple years, this product has been used experimentally over Guam (using Himawari-9 data), even though the product was trained using GOES-16 data. ABI and AHI channels are similar enough that a useful signal can occur. The animation shows the Himawari-9 Day Cloud Phase RGB (using Band 3, 5 and 13; those 3 bands plus Band 15 are used by the LightningCast algorithm to create probabilities) over northwestern Australia, showing both Darwin and the Tiwi Islands. The contours of the right show the likelihood of a lightning observation within the next 60 minutes. Note how some of the cirrus clouds (yellow in the RGB) have high probabilities whereas others — correctly — have low (usually in regions where the thick cirrus is dissipating). Careful inspection of the animations show that probabilities increase before deep convection is present. That is perhaps more easily seen in the toggle below of imagery at 0250 and 0350 UTC; at 0250 UTC probabilities are elevated in/around Darwin Australia (circled in cyan) with only a few glaciated clouds over land (i.e., clouds that are yellow in the RGB). At 0350 UTC, in contrast, well-developed convection is apparent (with continuing high probabilities) in the same region; lightning is probably occurring.

These LightningCast probability fields were computed using CSPP-Geo software; that software is nearing a beta release.

View only this post Read Less

Depending on the needed lead time, there are many ways to estimate the cloud cover on a given day/time. This includes climatology, long-range (global) models, shorter-range (regional) models and then satellite and other measurements. April 8th is a day of increased interest for cloud forecasts. The WFO at Fort Worth (TX) will be providing cloud forecasts, starting on March 29th.

Climatology

When one is weeks, months or years out, then climatology is the only option for a idea how cloudy a region might be.

Solar eclipse 2024 weather prospects: Q&A with an expert https://t.co/tZBSdefNBS pic.twitter.com/66lUQeQGip

— SPACE.com (@SPACEdotcom) March 16, 2024

The above figure (from this UW/CIMSS Satellite Blog post) is apparently “… all over the internet, showing the cloud climatology (or the study of climate) over the past 28 years compiled from GOES”. Or a loop of each April 8th geostationary image since 1979. Of course these aren’t forecast, just what has happened in the past. There are several other similar climatologies, based on MODIS or re-analyzed data.

Long-range (global) models

Once the event is within a week or so, long-range global NWP offer some guidance regarding locations of low pressure areas and some fronts, links of NOAA forecasts out to 5 and 8 days. A page (by Tomer Burg) with ensembles of NWP cloud forecasts. Cloud forecasts, by Fort Worth (TX) NWS will be starting on March 29th.

Shorter-range (regional) models

Once the target date is within a few days, higher resolution regional models will offer guidance. Most numerical prediction models do not include the effect of the reduced solar radiation associated with a total solar eclipse, but we know that the reduced surface heating can decrease clouds such as fair-weather cumulus. NOAA’s HRRR model (“total cloud cover”) is one that apparently will take this eclipse into account.

Each NWS WFO also offers a short-term cloud cover forecast (mouse over “sky cover” and then the times to the right). Or these experimental pages.

Satellites

The day of the event, one can look at many observations, those from satellite include NOAA’s GOES ABI, from many sources (NOAA, geosphere, (M1), slider, RealEarth and SSEC). Many links can be found to GOES animations. Which spectral bands or combinations to use depend on the time of day, etc.

Summary

Of course if one doesn’t see the Sun from the Earth during the eclipse due to heavy cloud cover, one can always see the moon’s shadow on the Earth from GOES and other satellites. including the eclipse from 2017.

The SUVI on the GOES also allows for routine images of the Sun.

Be safe.

H/T

Thanks to the UW-Madison, SSEC; SSEC Data Services and UW/AOS. Thanks also for the Eclipse Predictions by Fred Espenak, NASA’s GSFC. And thanks to those who have blogged regarding this 2024 event. Several of the images in this blog were made using McIDAS-X. Some models forecast out to 16 days, but the question is to what skill level for clouds. T. Schmit works for NOAA/NESDIS/STAR and is stationed in Madison, WI.

View only this post Read Less